Daily AI Briefing
Sunday, October 4, 2026
The editor's 6 picks from the 69 stories AIBriefs curated that day.
Get this briefing in your inbox every morning
China-Aligned TA419 Phishes US AI Policy Experts
Proofpoint attributes TA419 to credential phishing against AI experts at US think tanks, universities, and law firms since at least April 2025. A February 2026 email impersonating an Anthropic employee used the subject line "Request for Feedback on Military Integration of Claude."

OpenAI launches dots, always-on agents powered by GPT-6 Astra
Dots run on their own cloud computer, learn from feedback over time, and connect to over 4,000 apps through plugins. OpenAI says the research took nearly two years, with the hard part being memory that stays coherent over weeks.

Aleph Alpha releases Kolibri, a 78B open-weight German-English MoE model
Kolibri has 78.1B total parameters but activates only 3.46B (4.4%) per token, supports up to 1,048,576 tokens of context, and ships under Apache 2.0 with weights on Hugging Face. Aleph Alpha says it was trained on infrastructure in Germany and Finland and is built for sovereign work in regulated sectors.

Anthropic lobbied Vatican on AI consciousness, NYT reports
Per NYT reporting, Anthropic held months of NDA-secured meetings with religious scholars to argue its models may be conscious. Co-founder Christopher Olah reportedly threatened to pull out of Pope Leo XIV's encyclical launch after reading an advance copy of "Magnifica Humanitas," which states AI "do not understand what they produce."

Anthropic launches Claude Frontier Academy with $100M commitment
Anthropic is committing $100 million to train 10,000 Frontier Deployed Engineers by the end of 2027 through its new Claude Frontier Academy. First cohorts include engineers from Accenture, Bain, Capgemini, Deloitte, McKinsey, Morgan Stanley and Novo Nordisk.
Also that day
- LaunchOpenAI launches GPT-6 Astra, its first model to hit Critical cyber threshold
- LaunchOpenAI introduces GPT-6 Astra, billed as most intelligent and aligned model
- LaunchOpenAI rolls out GPT-6 Astra, its most advanced model yet
- AnalysisEpoch AI: Cost of a Given AI Performance Level Falls 47% Per Quarter
- LaunchAnthropic launches Cyber Mission with free OSS Scanner
- LaunchRunway introduces Solaris world interface model
- LaunchOpenAI's Agents API adds computer use, Bedrock managed agents
- EventFlorida seeks injunction to halt OpenAI frontier model development
- EventAI Alliance launches Project Tapestry for federated frontier models
- LaunchDeepSeek Harness v0.2 ships official desktop apps
- EventOpenAI and Anthropic probing tens of thousands of AI incidents
- AnalysisGoogle moves federated learning into TEEs for verifiable privacy
54 more stories that day — open the live feed.
Daily brief
Get tomorrow's AI brief in your inbox